Automated inverter noise separation in the interior of electric vehicles using adapted Kirsch-Compass kernel and Hough-Transformation

The increasing electrification of the powertrain of road vehicles creates new NVH challenges. In electric vehicles, the inverter noise is a dominant component of the overall interior noise. The typical high-frequency tonal noise shares often have a negative impact on the perceived sound quality of the vehicle. To improve the acoustic optimization, which is a complex and time-consuming process, a novel method for the automated separation of inverter noise shares in the vehicle interior has been developed and is presented in this paper. This method is based on image processing such as Kirsch-Compass filter and Hough-Transformation for line detection. It is designed to work even without the knowledge about component parameters such as the switching frequency and the number of pole pairs of the electric machine. The detection rate of the switching frequencies in the vehicle interior is 94%, and up to 86% for the inverter harmonics.